home sidereal designs

What internet marketers need to Know about proxy caching.



"The Desktop Global Marketer" (tm)

   A free on-line newsletter of Sidereal Designs, Inc.,
   for Internet Entrepreneurs, and those who are
   considering becoming one.
_____________________________________________________

                 February 2001

In this issue: "Why do you as an internet marketer
need to know about a technical issue called 'Proxy
caching'? Two reasons: it can keep your viewers from
seeing your updates and can interfere with your
hit counting."

_____________________________________________________

   "The Desktop Global Marketer" is free, and may be 
   re-published freely with permission. We encourage 
   you to give it to your friends.

   For subscription (or un-subscription) details,
   and other information, please see the end of the 
   newsletter.

   For any other purpose, please write to:
   	jamie(at)siderealdesigns.com
   Or visit us at:
   	http://siderealdesigns.com
_____________________________________________________ 

As an intrernet marketer you have two concerns that are
directly impacted by the widespread practice of proxy
caching. The first is that you want people to see the most
recent version of your web pages. The second is that you
want to know how many people are visiting your web
site. Proxy caching can confuse or defeat both of these
objectives.

Caching is a simple concept that can occur at several
levels, for example in your browser. When a Web page is
requested, it is saved to disk. If it is required again, the
disk copy is used instead of fetching the page again over
the internet. This technique is used in all modern browsers
such as Netscape Navigator and Internet Explorer. The
caching your browser does in your computer is set to expire
in a time period that you can set, anywhere from zero (no
caching), to once per session, or some time period. Most
people probably leave it at the browser default setting.

However, you may be being served cached pages by a middleman
as well. In this case a proxy cache, called an 'interception
proxy', receives users' requests for pages through a local
server such as their ISP (Internet Service Provider.) It then
provides cached copies itself instead of getting them direct
from the source. The local proxy server gets the page for the
first user requesting it, saves it on disk and forwards it to
the user. Subsequent requests from other users of the ISP's
cache get the saved copy, which is much faster and does not
consume Internet bandwidth. This saves the ISP a lot of
money.

An interception proxy thus functions as a middleman between
an end user and the web page the person is trying to
reach. The interception proxy pretends to be the destination
page, providing content or a network service faster to the
end user.

Interception proxies are used by AOL and other ISPs to
manage traffic from dial-up customers accessing the
Internet. In fact, 25% of the world's ISPs use interception
proxies.

Interception proxies violate the IP standard by breaking the
end-to-end nature of the communications technology.  The
proxies also alter communications without the knowledge or
approval of end users or destination page owners. On the
positive side, they decrease internet congestion.

What happens if you update your web page or add new
material? Well, your visitors may still see the old version
that's in cache. In theory a web site can stipulate a
time-to-live for a cached copy, or even 'no cache' in a
hidden part of the information it sends out with the page,
but it's up to the caching proxy server to honor that
request, and there have been many problems with certain very
large and popular ISPs failing to do so. Bandwidth costs
money and somehow 'no cache' requests seem to get
lost. Instances have been known of major daily newspaper
sites published on the web being six months out of date when
seen by users of AOL.

Still, it's a very good idea to be sure your site is
stipulating caching instructions for any pages you alter
frequently. At least the honest ISPs will honor it, and
you've done what you can. If it's a matter of serious
concern for your operation there are some technical tricks
you can pull to make things next to impossible to cache -
ask your webmaster.

But suppose you don't change your pages all that often, but
you would really would like to track the growth of visitors
to your site. Or, perhaps you've just spent a lot on some
costly advertising and you want to find out how much traffic
its getting you. So, you install some fancy statistical
hit-tracking software. Ooops! proxy caching is suddenly
giving you a big problem in interpreting those hit counts.

Why? Well, suppose that Joe at little-ISP.com looks at your
page. LittleISP.com doesn't do proxy caching, so Joe counts
as one hit. Then Fred at LittleISP does the same and you get
another hit. So far so good. However, when Linda at AOL
visits your site, she gets one count and the page gets
cached in a proxy server. Then Tom, and Mary, and Bill, and
... several thousand other people from AOL ... all visit
your site and they never really get there; they get the
cached copy at AOL. Your hit trackers never know about the
fact they've looked at your pages.

The two guys from Little counted twice as heavily as the
thousand people from AOL as far as your tracking software is
concerned, and your hits are several thousand short of the
number of people who actually looked at your page.

Unfortunately there is almost no way to guess how proxy
caching is affecting your numbers. If it's really important
you can resort to some of the technical tricks designed to
defeat cachers, but for most of us the best we can do is be
aware of the problem and use that awareness in interpreting
our statistics. What that means is treating them as relative
numbers. 

Don't ask about absolute figures, ask how much your hits are
going up or down over time, or by what percentage they
increased in the month following your ad campaign. 

When looking at advertiser's statistics, don't compare
apples and oranges: absolute hit numbers on web sites whose
viewers tend to be technically savvy shouldn't be compared
with absolute hit numbers from web sites whose viewers tend
to be novices coming from AOl; the former are much less
likely to be going through caches and as a result will count
proportionately more heavily.

This is only one of many ways marketing assumptions can be
distorted by the technological changes that are being
adopted, and forewarned is forearmed.

_____________________________________________________

To subscribe, send email to: 
	newsletter-request(at)siderealdesigns.com
and include the word   subscribe   as the only item in
the body of the letter.

To unsubscribe, send email to: 
	newsletter-request(at)siderealdesigns.com
and include the word   unsubscribe   as the only item in
the body of the letter.

If you have problems with either of these, write directly
to jamie(at)siderealdesigns.com for prompt attention from
a human.

If you would like to re-publish any of our newsletters,
at no cost, please contact jamie(at)siderealdesigns.com.

"Sidereal" is pronounced sy-DEER-ee-all, and means "of
or pertaining to the stars, the heavens, etc."


______________________________________________________________________________
Sidereal Designs, Inc. "Making The Web Simple."   http://siderealdesigns.com








Copyright © 1997-2000 by logo Sidereal Designs, Inc. All rights reserved.